5 ms·

Wonderful. I saw this paper and said to myself this is so neat I wish I could have done the work myself. Using statistics in this way to explain the quantifiab

Wonderful. I saw this paper and said to myself this is so neat I wish I could have done the work myself. Using statistics in this way to explain the quantifiable phenomena in language evolution makes for a milestone of the field.

thaumasiotes 5y ago
Ultradeep relationships between languages are to linguistics as perpetual motion machines are to physics. Everybody wants to let you know why their theory works better than everybody else's. There's a good page on zompist discussing the question of "just how likely is it that all these similar words are nothing but coincidence?"
- enkid 5y ago
  Seems like one of those times when statistics can tell you anything you want to hear.
- trhway 5y ago
  >"just how likely is it that all these similar words are nothing but coincidence?" like the word for mother in most of the languages is based on the sound consisting of "a" and "m" what the child makes while suckling https://www.proflowers.com/blog/56-different-ways-to-say-mom/ https://www.proflowers.com/blog/56-different-ways-to-say-mom...
  - Sharlin 5y ago
    "ma", "pa", and "da" are all among the first sounds a babbling baby typically learns to make. (As an aside, in my native language Finnish the most common word for mother, äiti, is notably not a "ma" word. It's apparently a loan from some Proto-Germanic word that has itself disappeared from extant Germanic languages. However, Finnish does have its own ancient "ma" word that means "mother": emä or emo, which is still in use when referred to non-human animal mothers, in poetry, and in a figurative sense in compound words such as emolevy "motherboard" or emäalus "mothership".)
- bradrn 5y ago
  Links: http://www.zompist.com/proto.html http://www.zompist.com/proto.html (with less statistics), http://www.zompist.com/chance.htm http://www.zompist.com/chance.htm (with more statistics).
mcswell 5y ago
Good statistics depends on good data. Theirs isn't. See the posts higher up for links to comments at Language Log.

thaumasiotes 5y ago

Ultradeep relationships between languages are to linguistics as perpetual motion machines are to physics. Everybody wants to let you know why their theory works better than everybody else's. There's a good page on zompist discussing the question of "just how likely is it that all these similar words are nothing but coincidence?"

enkid 5y ago
Seems like one of those times when statistics can tell you anything you want to hear.
trhway 5y ago
>"just how likely is it that all these similar words are nothing but coincidence?" like the word for mother in most of the languages is based on the sound consisting of "a" and "m" what the child makes while suckling https://www.proflowers.com/blog/56-different-ways-to-say-mom/ https://www.proflowers.com/blog/56-different-ways-to-say-mom...
- Sharlin 5y ago
  "ma", "pa", and "da" are all among the first sounds a babbling baby typically learns to make. (As an aside, in my native language Finnish the most common word for mother, äiti, is notably not a "ma" word. It's apparently a loan from some Proto-Germanic word that has itself disappeared from extant Germanic languages. However, Finnish does have its own ancient "ma" word that means "mother": emä or emo, which is still in use when referred to non-human animal mothers, in poetry, and in a figurative sense in compound words such as emolevy "motherboard" or emäalus "mothership".)
bradrn 5y ago
Links: http://www.zompist.com/proto.html http://www.zompist.com/proto.html (with less statistics), http://www.zompist.com/chance.htm http://www.zompist.com/chance.htm (with more statistics).

mcswell 5y ago

Good statistics depends on good data. Theirs isn't. See the posts higher up for links to comments at Language Log.